A novel hybrid architecture for classification of power quality disturbances

Power quality disturbances are introduced in the signals due to the increasing use of power electronic devices. To ensure reliability, security and adequate quality of power for consumption, the power quality disturbances need to be detected and classified accurately. This paper proposes an efficient algorithm for detecting and classifying the various power quality disturbances using a convolutional neural network (CNN) to extract various features from the input power signal which are then fed to the multi-class support vector classifier (MCSVC) to detect and classify the power quality disturbance events. It is observed from the simulation results and verified using an industrial dataset that the proposed model performs better than a normal convolutional neural network by approximately 10%. This work contributes to improving the quality of power delivered for industrial applications, making the operation of power systems economic, efficient and safe.